Possible CIFS bug - open one file, get another
Subsystems (2)
Hi folks.
I have what may be an in-kernel CIFS filesystem bug.
It seems that if I open a specific file, I get the contents of a different file back.
This may or may not have security implications.
Oddly, if I read the file with the same code in CPython 2.x, CPython 3.x or Pypy (recent trunk), I see the error. But if I run exactly the same code on Jython 2.5.x, it works fine.
Also, if I reboot, things work OK for a while.
And if I umount+mount, that works too.
I ran across the issue, because I have a backup program I've been writing for fun, and it saves the same timestamp inside a file, and in its filename. When I assert that the two are the same within some smallish tolerance, the assertion sometimes fails, as described above.
Is there a bugtracker somewhere that such an issue should go into?
Has this already been fixed in later kernels?
I'm on Linux Mint 12, kernel 3.0.0-12-generic.
Thanks!
A message that is not in the archive (sent off-list or not yet received): [email protected]
A network trace such as from wireshark or netmon (e.g. attached to a bug report) would help - as it would let us compare a working and failing scenario and see if the client is sending the wrong file name or the server is opening the wrong file in response to a request to open the correct file.
There is additional information at
http://wiki.samba.org/index.php/LinuxCIFS_troubleshooting
Quoted 36 lines · Dan Stromberg
Hi folks.
I have what may be an in-kernel CIFS filesystem bug.
It seems that if I open a specific file, I get the contents of a different file back.
This may or may not have security implications.
Oddly, if I read the file with the same code in CPython 2.x, CPython 3.x or Pypy (recent trunk), I see the error. But if I run exactly the same code on Jython 2.5.x, it works fine.
Also, if I reboot, things work OK for a while.
And if I umount+mount, that works too.
I ran across the issue, because I have a backup program I've been writing for fun, and it saves the same timestamp inside a file, and in its filename. When I assert that the two are the same within some smallish tolerance, the assertion sometimes fails, as described above.
Is there a bugtracker somewhere that such an issue should go into?
Has this already been fixed in later kernels?
I'm on Linux Mint 12, kernel 3.0.0-12-generic.
Thanks!
--
To unsubscribe from this list: send the line "unsubscribe linux-cifs" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
Signature · 3 lines
Thanks, Steve